When reading records from a database, the individually read database records may be treated as a stream. The stream of retrieved database records may be processed one after another in sequence such that a single record must be fully processed before the next record begins processing. Such stream data record processing requires the database application to maintain a long connection to the database. This consumes great amounts of server resources. Alternatively, such stream data record processing may require a client to read all of the desired data records prior to processing them. This consumes great amounts of client resources. Present systems may not provide an efficient stream processing interface to perform multiple parallel database queries.